Automatic ¹³C chemical shift reference correction for unassigned protein NMR spectra

Area of Science:

• Biophysical Chemistry
• Structural Biology
• Computational Chemistry

Background:

• Accurate chemical shift referencing is essential for protein Nuclear Magnetic Resonance (NMR) spectroscopy.
• Existing methods for correcting ¹³C referencing errors often require prior protein resonance assignment or 3D structure information, limiting their utility.
• ¹³C referencing errors can hinder or prevent crucial analyses like protein structure determination and dynamics.

Purpose of the Study:

• To develop a novel computational framework for detecting and correcting ¹³C chemical shift referencing errors in protein NMR data.
• To enable accurate referencing prior to protein resonance assignment, thereby facilitating downstream analyses.
• To lower the technical barrier for utilizing protein NMR for biomacromolecular studies.

Main Methods:

• Construction of a Bayesian probabilistic framework named Bayesian Model Optimized Reference Correction (BaMORC).
• Integration of BaMORC with an intra-peaklist grouping algorithm to create the Unassigned BaMORC method.
• Utilization of unassigned experimental peak lists and amino acid sequence for error correction.

Main Results:

• Unassigned BaMORC maintained ¹³C referencing errors within ±0.4 ppm for HN(CO)CACB-type peak lists.
• The base BaMORC method achieved ¹³C referencing error correction within ±0.45 ppm at a 90% confidence interval on a large unassigned dataset.
• Assigned BaMORC refined ¹³C referencing errors to within ±0.22 ppm at a 90% confidence interval with assigned chemical shifts.

Conclusions:

• The developed BaMORC framework effectively corrects ¹³C chemical shift referencing errors in protein NMR data.
• Unassigned BaMORC provides accurate referencing early in the analysis pipeline, before resonance assignment.
• These methods enhance the accessibility and reliability of protein NMR analysis for a wider range of researchers.